Traditional probabilistic approaches consider MAD (multiple-wavelength anomalous-dispersion) data as a special MIR (multiple isomorphous replacement) case. The rigorous method of the joint probability distribution functions has been applied to solve the phase problem, with the assumption that the anomalous scatterers' substructure is a priori known. The probabilistic approach is able to handle measurement errors: it has been applied to symmetry-restricted phases and provides simple and efficient formulas.